Plate layout

Shows the layout of samples, controls (e.g., negative controls (NC), Inter-plate controls (IPC) and special wells if any (e.g., Sample controls (SC)) on the 96 well assay plate.

NOTE: Plate ID variable had missing or duplicate entries so plate IDs were automatically assigned.

Plate_01

Plate_01: Covid_Example.xml
1 2 3 4 5 6 7 8 9 10 11 12
A
B NC rep01 NC rep02 NC rep03 IPC rep01 IPC rep02 IPC rep03 Aged matched control Z001-1 d1 Aged matched control Z002-1 d1 Aged matched control Z003-1 d1 Aged matched control Z004-1 d1
C mild RS42 d1 mild RS42 d3 mild RS42 d5 mild RS42 d7 mild RS42 6 d21 mild RS42 d14 mild RS53 d4 mild RS53 d8 mild RS53 d16 mild RS53 d21
D control RS43(H98) d1 control RS43(H98) d3 control RS43(H98) d5 control RS43(H98) d7 control RS43(H98) d14 control RS43(H98) d21 HC H1 d1 HC H4 d7 HC H4 d1 Aged matched control Z005-1 d1
E mild RS47 d1 mild RS47 d3 mild RS47 d5 mild RS47 d8 mild RS47 d14 mild RS47 d21 long COVID VB01 d1 long COVID VB02 d1 long COVID VB03 d1 long COVID VB04 d1
F mild RS48 d3 mild RS48 d6 mild RS48 d8 mild RS48 d22 long cov control UK01 long cov control UK03 long COVID UK04 long cov control UK05 long COVID UK06 long COVID UK07
G mild RS49 d1 mild RS49 d3 mild RS49 d6 mild RS49 d8 mild RS49 d16 mild RS49 d22 long COVID UK09 long cov control UK10 IPC rep04 NC rep04
H

Plate_02

Plate_02: Example_01.xml
1 2 3 4 5 6 7 8 9 10 11 12
A A 01 C001 P3 A 02 C044 P1 A 03 C028 P3 A 04 C032 P1 A 05 C026 P2 A 06 C015 P1 A 07 C001 P2 A 08 C022 P2 A 09 C008 P2 A 10 C037 P2 A 11 C001 P4 A 12 SC rep01
B B 01 C035 P1 B 02 C006 P1 B 03 C031 P2 B 04 C021 P3 B 05 C012 P1 B 06 C039 P4 B 07 C021 P2 B 08 C028 P2 B 09 C032 P2 B 10 C011 P2 B 11 C036 P2 B 12 SC rep02
C C 01 C043 P2 C 02 C045 P3 C 03 C027 P2 C 04 C010 P3 C 05 C024 P4 C 06 C024 P1 C 07 C003 P2 C 08 C031 P1 C 09 C011 P1 C 10 C039 P1 C 11 C033 P1 C 12 IPC rep01
D D 01 C034 P2 D 02 C009 P4 D 03 C008 P4 D 04 C044 P2 D 05 C004 P1 D 06 C021 P1 D 07 C039 P2 D 08 C031 P4 D 09 C027 P1 D 10 C036 P1 D 11 C015 P2 D 12 IPC rep02
E E 01 C045 P1 E 02 C014 P2 E 03 C041 P1 E 04 C020 P3 E 05 C028 P1 E 06 C029 P2 E 07 C035 P2 E 08 C040 P4 E 09 C020 P4 E 10 C002 P1 E 11 C001 P1 E 12 IPC rep03
F F 01 C043 P1 F 02 C006 P2 F 03 C024 P3 F 04 C040 P3 F 05 C020 P1 F 06 C045 P4 F 07 C031 P3 F 08 C010 P1 F 09 C010 P2 F 10 C034 P1 F 11 C026 P1 F 12 NC rep01
G G 01 C042 P1 G 02 C014 P1 G 03 C020 P2 G 04 C009 P2 G 05 C038 p1 G 06 C005 P2 G 07 C008 P1 G 08 C037 P1 G 09 C009 P3 G 10 C044 P4 G 11 C041 P2 G 12 NC rep02
H H 01 C017 P1 H 02 C033 P2 H 03 C046 P1 H 04 C008 P3 H 05 C003 P1 H 06 C026 P3 H 07 C038 P1 H 08 C012 P2 H 09 C045 P2 H 10 C022 P1 H 11 C014 P3 H 12 NC rep03

Read summary

Summary of the number and classification of reads for the assay run

  • Parseable - Valid next generation sequencing (NGS) read (assay signal or background)
  • Parseable Match - Valid NGS read (assay signal)
  • Parseable Non-match - Vaild NGS read (assay background)
  • Unparseable - Sequencing background
  • Total wells - Total number of samples + controls on plate
  • Total targets - Total number of assay targets (including internal controls)
  • # of zeros (%) - Number and percentage of targets * samples which had zero sequencing reads (no assay signal)


Internal controls (IC)

Each sample of the assay is spiked with the same concentration of at least one internal control (IC). This is used both as part of a well normalization procedure and as a method for assessing the uniformity of the assay run. SD: standard deviation. CV: coefficient of variation

Internal Control (IC) summary - mCherry


Inter-plate controls (IPC)

IPC samples are pooled plasma controls used both to normalize samples between experiments and to assess performance (e.g., precision) of the assay.

Note: %CV in this table does not exclude values below LOD. See Coefficient of variation section for %CV (before and after normalization) that excludes values below LOD.

IPC summary


Negative controls (NC)

NC wells are reactions where no sample input is provided (i.e., buffer only). These are used to assess Limit of Detection (LOD) for each target assay and to assess run quality (e.g., background levels).

NC summary


Heatmaps

Heatmaps show the percent relative to plate median for log2(total counts) and the specified internal control(s).

Plate_01

Plate_02


Quality control

Plate QC

Plate-specific QC criteria:

  • IC CV - Coefficient of variation of internal control Parseable Matching reads across all wells (Maximum threshold = 25%)
  • IPC CV - Coefficient of variation of Parseable Matching reads across all IPCs (Maximum threshold = 25%)
  • IPC Target CV - Median coefficient of variation of all IPC targets (Maximum threshold = 10%)
  • Run Detectability - Percentage of targets that are detectable (target is considered detectable if > 50% of samples are above LOD) (Minimum threshold = 90%)
  • Reads - Minimum number of Parseable Matching reads for the run (Minimum threshold = 100,000,000)

QC criteria are assessed either on unnormalized (raw) or IPC-normalized (IPC) data.

Number of warnings: 0

Combined Plate QC Summary

Sample QC

Sample QC summary tables

Sample-specific QC criteria:

Number (%) of samples that have warnings for each QC criterion (excludes IPC, SC & NC samples)

Wells with warnings: Detectability

Wells with warnings: Reads

Wells with warnings: IC Median

Sample QC plots

Target QC

Target QC summary tables

Target-specific QC criteria:

  • Target Signal - Percentage of samples with raw reads of at least 200 (Minimum threshold = 50%)
  • CV Normalized Reads - Target-specific SC intra-plate coefficient of variation of normalized reads (Maximum threshold = 30%)
  • Target Detectability - Percentage of samples with signal above the limit of detection (Minimum threshold = 50%)

Number of Targets with warnings: 206

Number (%) of targets that have warnings for each QC criterion

Target Signal

CV Normalized Reads

Target Detectability

Target QC plots


Detectability

Target detectability is the percentage of samples that are above the limit of detection for that target. A target is considered “detectable” if it is above limit of detection in greater than 50% of samples. “Overall” detectability is the overall percentage of samples across all plates that are above LOD for a given target. Target detectability reported here excludes IPC, SC, & NC samples.

Detectability summary

Plate_01

Plate_02

Overall

Target detectability table

Plate_01

Plate_02

Overall

Target detectability boxplot

Plot shows target IC-IPC-normalized count distributions relative to LOD. For each target, log2(LOD) was subtracted from log2(count) for IC-normalized data. Reverse-curve quantified targets are denoted by a “*” and are plotted as forward-curve targets for compatibility


Inter-plate CV summary

%CV was calculated for each target (excluding internal controls) for the SCs. Only values above LOD were used to calculate %CV. Some targets may have missing CV due to too many values below LOD.

Inter-plate CV table

IPC inter-plate CV


Sample boxplots

Sample boxplots show distributions of the log2 counts for each sample, unnormalized and after IC + IPC normalization.

Plate_01

Plate_02


Sample correlation

Plot shows the Pearson correlation between samples. Hierarchical clustering is done using complete linkage.

Plate_01

Plate_02


Sample and target clustering

Hierarchical clustering of both samples and targets was done using a Euclidean distance metric and complete linkage. Data is log2 and IC-IPC normalized; targets are centered and scaled.

Plate_01

Plate_02


Sample PCA


Plate effect assessment

Plate effect: Plate QC summary

A plate will have a plate effect QC warning if 10% or more of targets are significant. Significant targets have a significant global ANOVA F-test (unadjusted p < 0.01) and significant pairwise tests (Tukey-adjusted p < 0.01) in comparisons with 50% or more of the other plates.

Number of warnings: 2

Plate effect QC summary table


Plate effect: Target QC summary

A target will have a QC warning if plate ICC > 10%, significant F-test (unadjusted p < 0.01) and significant pairwise tests (Tukey-adjusted p < 0.01) for at least one plate.

155 targets have a plate effect QC warning.

  • Fstat - Plate effect ANOVA F-statistic. Targets are sorted in order of F statistic.
  • Ftest pval - ANOVA F-test p-value. P-values are unadjusted.
  • ICC plate - Inter-class correlation, the percentage of variability attributable to plate.
  • % sig pairwise - Percent of plates that show significant pairwise differences in mean with at least 50% of other plates
  • % plates low IPC - Percent of plates that have median IPC reads < 200.
  • % plates low sample reads - Percent of plates that have median sample reads < 200.

Targets with plate effect QC warning


Plate effect ICC scatterplot

A linear mixed effect model was fit for each target with plate ID as random effect. The plate-related intra-class correlation is the ratio of the estimated plate-related variance to the sum of plate and residual variance. This is multiplied by 100 and can be interpreted as the percent of variance attributable to plate. Target is highlighted if ICC > 10%.

155 targets have plate ICC > 10%.

Target boxplots

Boxplots are shown for targets with plate effect QC warning. A target will have a QC warning if plate ICC > 10%, global ANOVA F-test is significant (unadjusted p < 0.01) and target has significant pairwise tests for at least one plate (Tukey-adjusted p < 0.01 with at least 50% of other plates). Red asterisks above boxplots indicate plates showing significant pairwise test with 50% or more plates.

155 targets have a plate effect QC warning.

Plots for the top 30 targets based on plate ICC are shown below.

Target means by plate heatmap

Sample mean NPQ for each target is shown by plate in the heatmap. Target values are centered and scaled. Control samples were excluded from mean calculations.

PCA by run

PCA was performed on the sample mean by target and plate matrix shown in the heatmap above.


Sample PCA

PCA shown below was performed on the samples only, excluding any control samples. Samples are colored according to plate.

Pairs plots

Pairs plots show the first 10 PCs plotted against each other. Any clustering by plate in the plots indicates potential plate effect(s).

PC correlations with metadata

Sample metadata variables including plate and sample matrix, if provided, were correlated against the principal component scores for the first 10 PCs. P-values are adjusted using the Bonferroni correction. Significance scale for adjusted p-values: * < 0.05, ** < 0.01, *** < 0.001.


Plate effect summary table

  • n % sig targets - Number and percent of targets that show significant pairwise difference in sample mean for that plate versus 50% or more of the other plates. Significant targets have a significant ANOVA F-test (unadjusted p < 0.01) and significant pairwise differences have Tukey-adjusted p < 0.01.
  • sig targets - Targets with significant F-test and showing significant pairwise difference for at least 50% of plates.
  • n % low IPC reads - Number and percent of targets with median IPC raw reads below 200.
  • low IPC reads targets - Targets with median IPC raw reads below 200.
  • n % low sample reads - Number and percent of targets with median sample raw reads below 200.
  • low sample reads targets - Targets with median sample raw reads below 200.

Number of warnings: 0

Plate effect statistics by plate